EAST AFRICAN REPTILES AND AMPHIBIANS 49 



presence or absence of a mucro is not a character of specific impor- 

 tance, but appears to be affected by the nature of the environment. 



I suggest that typical West African Agama agama agama is mainly 

 an arboreal lizard characterized by a less depressed body, exceedingly 

 strongly keeled and mucronate scales, and a strongly raised ridge or 

 keel to the tail. As this lizard extends eastward through the forest 

 belt to Uganda, a reduction in the strength of the keels and mucros 

 has taken place and the caudal keel is generally absent or so low as 

 to be inconspicuous. Some name is probably available for the Cen- 

 tral African form, but in this paper I am not distinguishing it from 

 forma typica; at best it is only an intermediate between the western 

 and eastern lizards. 



I regard A. a. 'planice'ps of Angola and south West Africa as the 

 savannah representative of the typical forest agama; according to 

 Boulenger the range of the typical form extends southward to Angola. 

 Its southwestern race is characteristized by a more depressed body, 

 strongly keeled scales, with or without a mucro, a caudal keel in the 

 male. In 1919 Boulenger recorded its occurrence at Stanleyville, 

 Belgian Congo; in 1900 Roux lists one from Uganda wdiile Tornier 

 in 1897 refers material from several localities in Tanganyika Territory 

 to -planice'ps. In 1907 Lonnberg identified two lizards from the Meru 

 steppe with this species. I think that these East African records are 

 probably referable to some of the undermentioned races and are 

 subspecifically distinct from the Angolan lizard. 



In 1896 Boulenger described a lizard from southeast of Lake 

 Rudolph as closely related to planiceps and named it lionotus, its close 

 affinities can best be expressed by calling it ^. a. lionotus. It was 

 said to differ from planiceps in the larger spines on the side of the 

 head and neck, which nearly equaled the tympanic diameter, and by 

 the very feeble keeled dorsal scales. Incidentally the head is said 

 to be as long as broad. I feel sure that there is some error in this 

 statement. The life-size illustration shows the head longer than 

 broad, as is the case with all agamas of this group. The size of the 

 head and neck spines is a character of no significance among these 

 races, varying much among individuals from the same locality. In 

 the Museum of Comparative Zoology is an almost, or actual, topotype 

 of lionotus from the Guaso Nyiro district of Kenya Colony. This 

 male (M. C. Z. 8202), except for the slightly shorter spines on head 

 and neck, agrees so closely with the figure, description, and coloring 

 of the type that it leaves no doubt as to its identity. There is a female 

 (M. C. Z. 7993) from the same locality and a male (M. C. Z. 8204) 

 from the Ithanga hills also in the collection. 



In 1897 Werner described four agamas collected by Schweinfurth 

 as A. flavicauda but gave no locality. We know, however, that 



